Electronic voting systems aim to modernize electoral processes by improving efficiency, accuracy, and accessibility. However, concerns about transparency, security, and voter anonymity have hindered their widespread adoption—especially in countries where traditional paper-based systems remain deeply rooted. This article presents an enhanced version of the OTP-Vote model, a cryptographically secure system based on One-Time Pad encryption and multi-channel data storage. Two major contributions address common criticisms which are the integration of End-to-End Verifiability (E2EV) and the inclusion of accessible audit mechanisms. E2EV enables voters and the public to verify vote integrity without compromising secrecy, while audit enhancements offer traceability and support for manual verification. Mathematical modeling and simulations confirm a negligible risk of voter identification. Machine learning techniques were applied during the simulation phases to improve scenario analysis and optimize key parameters. These improvements aim to enhance transparency and foster public trust, supporting the safe adoption of electronic voting in democratic systems.